Sake Hoekstra schrijft: "Op www.m3dzone.com. Misschien wel belangrijk om te posten, want niets is erger dan een PC die vastloopt":

We krijgen steeds meer mailtjes van mensen die problemen hebben met hun Savage 4 videokaart, als bijv. de Diamond Stealth III. De meest voorkomende problemen zijn toch wel het vasthangen van het systeem bij het spelen van 3D Games. Gelukkig is er nu wel een oplossing, want er is nu RebuM's Savage4 driver voor Super7 chipsets v1.0b. Deze driver is fijn afgesteld door RebuM van the Union of Unreal Clans Europe. Deze driver zorgt voor vrijwel 100% stabiliteit bij het spelen van games die direct3d of metal gebruiken. Ikzelf had dit probleem dus met Unreal en Quake en met deze nieuwe drivers is dit opgelost! Download ze dus nu vanaf onze website.

De file (gebaseerd op de 8.10.07 drivers van S3) kun je hier neerhalen (1,4Mb).

Hilko Rigterink schrijft: "Eindelijk! Na lang wachten komt dan op 22 oktober het 2e deel uit van GTA, oftewel Grand Theft Auto:":

The bad boys of gaming, DMA Design, have returned and are jamming the hammer down on the boundaries of political correctness once again with their overhead action game, GTA2. The PC and PlayStation sequel to the classic no-holds-barred crime simulator has gone gold and is anticipated to be in stores Oct. 22.

Misschien bouwen ze niet de meest vooruitstrevende 3D accelerators van dit moment, maar enig gevoel voor humor heeft men er wel... Deze ansichtkaart ontving AGN 3D van 3dfx :

3dfx passed along the following teaser postcard, to give us an idea of what is coming up the channel for what most likely will be a Comdex showing. I would assume that we are going to see the Voodoo4 in action, but then again does the Voodoo4 need toilet paper in large rolls? Or maybe they plan on wiping up the competition with their massive rolls of power?

Bunny (wederom zo'n rondstuiterende overklokker uit Japan) heeft benchmarks gepost van een Athlon 600 systeem op 810MHz met 125MHz SDRAM (3,25*2*125). Als koeling werd een water-cooled unit van Kendon gebruikt.

Ik kwam bij Ace's Hardware een posting tegen over het effect van de AMD AGP miniport drivers op de video performance van een Athlon systeem. Zonder de drivers kan de framerate met 50% naar beneden duikelen:

Be sure to check that you/ your reseller installed the AGP miniport (now version 4.43) when you install windows 98 on an Athlon system. I have seen with my own eyes, that some resellers do not install this chipset driver and just let Windows 98 install the drivers. Several readers have confirmed this. If no special AGP port is installed, windows 98 does not even install an AGP chipset driver. If you can find "AGP to PCI host bridge" among your system devices, everything is ok. If You can't, you now know why your system "crawls". Especially with the Geforce 256, the AGP miniport is extremely important; here are the results of Quake 3, with the 3D Prophet (GPU : nVidia Geforce 256), 3.35 OpenGL drivers (T&L acceleration).

[...] No AGP mini port = abysmal performance ! Running Indy3D, the PIIIB was even twice as fast as the crippled Athlon system ! To be perfectly clear: The AMD 750 chipset has no problems at all with Videocards and AGP. It just requires specific chipset drivers.

Check Ace's Hardware voor benchmarks.

Femme schrijft: "Ace's Hardware heeft een artikeltje in elkaar gedraaid waarin ze uitleggen hoe je een Apache webserver (in dit geval onder Solaris) kunt installeren en configureren":

Interested in becoming a part of the most exciting and ever expanding segment of the Internet today, the web? If so, you may also be interested in using your own webserver. In this article, I'll take you through the initial setup and configuration of such a server. Along the way, we'll examine our needs, and determine exactly what we need to get the job done.

This article is more of a "hands-on" How-To guide more than anything else, if you're interested strictly in web server performance, you'll be best suited to read my article, Building a Better Web Server, which examines different platforms and server models.

Since the free Apache web server accounts for roughly 50% of all web servers and Sun claims the majority of the hardware side of things, I'll be overviewing such a configuration. In fact, for convenience above all, the server I'm detailing is none other than Ace's Hardware's very own web server.

Check het artikel hier.

BDW: deze posting fungeert als voorbeeld van de werking van ons nieuws nieuws submit formulier, dus niet denken dat ik in mezelf loop te lullen .

Hardware guru Tom heeft naar wat high-poly games en demo's gekeken die speciaal gemaakt zijn voor de T&L engine van de GeForce 256. In z'n artikel staan genoeg plaatjes van dergelijke titels, zoals 'Dagoth Moor Zoological Gardens'. Om het zaakie compleet te maken heeft Tom er ook nog wat benchmarks doorheen gemixt. Lees het hele verhaaltje hier.

Well, things haven't really changed significantly since our first review of GeForce. The T&L-engine is a great feature as long as an application makes use of it and GeForce was able to show its muscles quite well in Dagoth Moor Zoological Gardens. However, the performance of the rendering pipeline really doesn't knock me off my feet, especially when single data rate memory is used on the card. It constrains the rendering pipeline with its low memory bandwidth even more.

Dragon Moor Zoological Gardens however did knock me off my feet. I want to see more games like this. I want tons of polygons and 3D-worlds that are worth being looked at for a change. Dagoth Moor Zoological Gardens is definitely the way to go. Now we only need the hardware for it. GeForce, the pioneer of all 3D-chips with integrated T&L, is a step in the right direction. Games like Dagoth Moor Zoological Gardens will soon make an integrated T&L-engine a common requirement. NVIDIA is with GeForce256 right where the rest of the 3D-industry wants to be. Future 3D-chips without T&L will probably have a very hard time.

Graig Barrett, hoogste pipo bij Intel, heeft op de Gartner ITExpo door laten schemeren dat het volgens hem nog 10 jaar kan duren voordat 64-bit zachtgoed & processors volledig naar de desktop ingeburgerd zijn. Dit terwijl men eerder gezegd heeft dat de McKinley (opvolger van Itanium) al in 2002 richting mainstream moet zakken. The Register heeft er het volgende over op te merken:

He did have some ominous words to say about the move to 64-bit, suggesting that it could be five to ten years before it would be the norm on the desktop; he compared this to the ten years between the introduction of the 80386 and the significant development of 32-bit desktop applications. If it's really going to take ten years, then we can maybe start wondering about Intel's roadmapping. The company officially reckons McKinley, the iteration after Merced, will be the 64-bit Intel chip that moves into the mainstream, but that's hardly a ten years from now product.

Intel does intend to keep plugging away at 32-bit alongside 64-bit for a few years, so Barrett could be signalling that this plan will be extended. Or alternatively, we could bear in mind that Intel was happily selling 386 chips for several years before the software started to catch up.

But there were reasons for the slow development of the software that probably don't apply these days. At the 386 introduction Microsoft and IBM were in the driving seat (and incidentally still squabble over whose fault it was they originally aimed OS/2 at the 286). These days we've got Linux and several flavours of Unix aimed at IA-64, so take-up is likely to be a lot faster. But that's software, and a problem for Microsoft - so Barrett should worry.

Volgens The Register zal Intel eind deze maand de i840 Carmel chipset introduceren. Tegelijkertijd wordt er een rond deze chipset gebouwde server plank op de markt gepoot:

The board, dubbed the OR840, is expected to cost around $250 for the boxed version and will support up to two Intel CPUs.

It will come with support for up to 2Gb of memory, and four 168-pin RIMM (Rambus) sockets, supporting speeds of 600MHz and 800MHz. That indicates that Intel's problem with Rambus memory may be confined to the i820 Camino chipset only.

[...] The OR840 has built in support for floppy drive, ultra DMA and integrated Fast Ethernet 10/100 Base TX. Video, which is not integrated, has a 266MHz data transfer rate, while the board has two USB ports, one 32-bit PCI slot, and five other bus slots. Bus slot four uses AGP 4x.

The Register zegt te beschikken over de introductieprijzen van de twee Coppermine mobiles, die eind deze maand tegelijkertijd met de desktop en Xeon versies gereleased zullen worden:

The 450MHz Pentium III/mobiles will cost $360 each when bought in boxes of 10, and $350 each when bought in boxes of 100.

The 500MHz Coppermine Pentium III/mobile will command a price of $545 when bought in units of 10, and $535 when bought in units of 100. Both come with 256K cache.

The introduction of these parts spells the real beginning of a shift away from .25 micron process technology to a situation where Intel's processors will eventually all have a smaller die size, and run cooler and faster.

The Register heeft een nieuwsartikel neer geplakt met meer info over de USB 2.0 specs die gisteren door de USB Promotor Group bekend werden gemaakt (zie ook deze nieuwsposting):

The USB Promoter Group clarified its stance on IEEE1394 (aka FireWire and iLink): USB 2.0 will become the be all and end all of PC connectivity and 1394 will continue to exist in the consumer electronics space, connecting digital VCRs to digital TVs and the like. In the USBPG worldview, the only PCs that require 1394 will be those that need to connect to such devices -- all other high bandwidth devices, such as fast hard drives, will clearly be hooked up via USB 2.0.

As the Technical Introduction to USB 2.0 puts it, even "high bandwidth interfaces such as SCSI adapters may no longer be required". So where we once had 1394 replacing SCSI while USB replaced the old serial and parallel ports, we now have USB 2.0 replacing everything.

USB 2.0 is unlikely to supersede 1394, though, since it retains the need for a hub -- unlike 1394, there's no peer-to-peer operation; data can only from device to device via the hub. In fact, hubs will take on a greater role with USB 2.0 since they will have to arbitrate data traffic flowing at 480Mbps and the original 12MBps and 1.5Mbp

Jaja, je gelooft het bijna niet maar er zijn nu ook al artikelen over artikelen
Bill van g256.com heeft een overzichtje van previews van de Geforce256. Zo kun je zien waar je welke info kunt halen en welke je niet hoeft te lezen. Volgens Bill heeft Anand de beste review en kun je het beste richting Planetriva fietsen voor screenshots/overclocking.

Over Anand:

By Far the best review, not only for benchmarks, but for everything else except screenshots. He covers many cpus, and many different benchmarks, If you only want to read one review this is it. Very easy to navigate this monster review! Covers overclocking, They even review with the AMD k-6! [break]Over Tom zijn ze wat minder te spreken:[/break] He's pimping this for all its worth. Tom splits the benchmarks on many pages, each graph taking almost a whole page size. His conclusions are ok but not very original. I'm not out to trash Tom but Jeeze! [break]En even verder:[/break] I really wanted to put Tom's last but I couldnt with good concience.

Check hier voor het hele verhaal.

Dit keer geen wazige Japanner die al solderend weer wat nieuws ontdekt, maar een Australiër. Hij krijgt het voor elkaar om voltages van 1,8 tot 3,5 volt op elk bx moederbord beschikbaar te maken. De techniek van het aftapen van bepaalde pinnen om het voltage op te krikken kenden we al, maar dit figuur pakt het anders aan:

While trying this I came up with an easy method of allowing all core voltages 1.80-3.5 on any slot one board. To understand how this works you must know how the pins select the voltage. Voltage is selected by pins VID0 - VID4. These pins can either be open (not connected to anything) or earthed (connected to Vss - motherboard earth). The processor VID pins are connected to Vss or left open when the processor is made, a different combintion for each voltage, allowing it to tell the MB what voltage it needs, taping allows us to open an earthed pin thus giving the alternative voltages 2.2, 2.4 etc.
OK so the MB reads the processor VID pins and sets the voltage, why not bypass the CPU pins completely and fool to MB into giving us any core voltage. This is more simple than it sounds.

Vervolgens volgt een heel stappenplan. Het handige is dat hij de boel aansluit op een blokje dipswitches. Zo kun je heel simpel het gewenste voltage instellen. Uiteraard is enige handigheid met de soldeerbout vereist . Volg deze link voor het hele artikeltje en wat foto's

Bijl heeft een begin gemaakt aan een BP6 users FAQ. Er zijn nogal wat vragen over dat moederbord en veel ervan worden steeds maar opnieuw gesteld...

De FAQ is nog lang niet af (dus zeg niet "hee waarom is die FAQ niet af", want dat heb ik net gezegd ). De eerste aanzet staat in ieder geval op www.bp6.com
De bedoeling is dat er aanvullingen van users komen, maar dat wil nog niet echt vlotten

Hier heb je alvast een paar voorbeelden

Q: Do I have to cool my chipset (the one with the green heatsink) extra? A: In SMP-mode the chipset gets hotter than normal. It is a good idea to apply heatsink compound. Q: But the heatsink on my chipset isn't even hot... A: That's because the heatsink isn't doing its job: after you have applied the heatsink compound you will notice that the heatsink is warmer, it's now conducting heat better Q: I've connected a case fan to the FAN3-connector but I cannot read the RPM. What's wrong? A: Nothing. RTFM. This connector is not equipped with the RPM feature. Q: I've got a exclamation mark in W2K at my USB device I want to get rid of... A: Set Your MPS to 1.1 when using Win2K to solve USB Root Hub problems. MPS 1.4 seems to cause the USB ports to not function.

JC heeft cool nieuws over geruchten dat AMD een tweede Athlon chipset in de planning heeft. De AMD760 zal naast AGP4x support waarschijnlijk ook ondersteuning voor PC2100 DDR SDRAM meebrengen:

A very anonymous fellow who has recently proven himself of great worth to me subtly hinted (okay, he banged it over my head) that in addition to VIA's KX133, AMD will be adding another Slot-A chipset to the mix. Specifically, AMD is supposedly going to unveil the AMD760 chipset, which has AGP4x and I hope some nonmarketing advantages, like DDR SDRAM support (mmmm...maybe DDR VCDRAM!), towards the end of the next quarter (Q1). Some sort of SMP chipset is expected at the end of Q2. I place that second bit under a microscope, as we've heard of rumours of dual chipsets before that point (like from Kryotech), so AMD's first SMP chipset will hopefully be unveiled before then! :O

Techweb bericht dat Intel in het vierde kwartaal, waarschijnlijk november, een tweede poging zal ondernemen om de i820 Camino chipset in de markt te planten. Als deze info klopt dan is de delay dus minder dan eerder werd aangenomen (~rond de 3 maand):

Intel's twice-delayed Camino PC chip set will ship in the fourth quarter, and most likely in November, according toindustry sources that were briefed by Intel late yesterday.

Intel, Santa Clara, Calif., professed no knowledge of the meeting, which several sources said was conducted in advance of the company's quarterly earnings conference call set for Tuesday afternoon.

"There have been no public updates on Camino," said a company spokesman, who added that platform integration work continues and no new launch date has been scheduled.

That may be true, but sources said they were specifically told that Intel would roll out the Camino, or Intel 820 as it is officially known, in the fourth quarter. They said the company also indicated that it would be able to ship the product at that time, though only in small quantities. Earlier this week, Intel publicly denied the Camino would be released in the first quarter of 2000, contradicting published reports.

Over de i840 Carmel server chipset werd niets gezegd, maar verwacht wordt dat deze op 25 oktober tesamen met de PIII Coppermine geïntroduceerd zal worden. Zie Techweb voor het complete nieuwsartikel.

3Com heeft details bekend gemaakt van een nieuwe technologie voor draadloze communicatie, waarmee snelheden van 2Mbit per seconde gehaald kunnen worden. Volgens Newsbytes is dat 4 keer zo snel als een 56K6 modem (laatste keer dat ik 2000/56 deelde lag de uitkomst nog rond de 40 ):

Before the end of next year, 3Com Corp. [NASDAQ: COMS] will unveil technology that allows Internet users to connect wirelessly, via laptop, at speeds of up to two megabits per second (Mbps) - more than four times faster than a standard 56 Kbps dial-up connection - 3Com executives announced at a telecommunications conference in Geneva today.

With typical wireless connection speeds hovering at around 14 Kbps, the new technology represents a massive leap in the realm of wireless connectivity, 3Com spokesperson Mark Bingham contended.

3Com is developing the technology as part of a newly announced three-pronged wireless strategy aimed at providing wireless connections to consumers at home, at work and "on the move."

[...] Home consumers will be able to use the wireless LAN technology to connect to the Internet, play networked games and connect multiple computers "without punching holes in the walls and laying Ethernet cords all over the house," he said.

Ars Technica heeft een review van de el cheapo Apex ATX Full Tower op het net gekwakt. Geen über tower zoals m'n SuperMicro SC760-A , maar wel een redelijk kastje gezien de lage prijs ($70):

As a whole I'm fairly impressed with this full tower offering from Apex. It's no SuperMicro SC-750a, but at one-half the price it forces you to think, "are the differences really worth all that money?" It's definitely quieter than a 750a, but that comes at the expense of less internal air flow cooling. It offers more available drive bays that the 750a, but at the expense of some open space. The case is not quite as "tank-solid" as the 750a, however, it's stable enough given the fact that you're not going to beat on it with a baseball bat.

But enough comparisons. If you're looking for a quality full tower case at a mid-tower price, then Apex's ATX Full Tower case is worth your time and attention. The case is generic to some extent, but it's very functional and really gets the job done. I'm using it for my main system, and as you can see above my praises smother my complaints.

Sharky Extreme heeft een update gepleegd aan hun Weekly CPU prices column, met o.a. info over de beschikbaarheid van de Athlon 700:

So far we haven't been able to locate an Athlon 700 CPU for sale that wasn't locked into a fully configured system, but reports are beginning to trickle in that indicate the OEM white box models have begun to ship out of AMD's warehouses.

These initial shipments will be available within the next seven to ten days and will cost in the area of $800 to $900 each. Wide spread Athlon 700 CPU distribution will occur by the last week of October as volumes ramp up, and November should find most AMD resellers stocking the high-speed chip without a problem.

FPS3D.com heeft een review gepost van de Quantum Vuurbal KA en vergelijkt de schijf met een 5400rpm Maxtor DiamondMax 6800. Ondanks z'n hogere rotatiesnelheid blijkt de Fireball niet veel sneller te zijn dan de Maxtor harddisk:

The Quantum drive does hang out at the top of the crop in terms of performance, but it's really hard to justify buying it, since such "slower" drives as the DiamondMax 6800 can almost knock it to the floor in the benchmarks. Not to mention without creating noise pollution.

If anything, buy this drive for it's stability. It has the most shock protection features of any hard drive in it's price range. Also, it can sustain perfect functionality under fire of overclocking. If you need a hard drive that will protect your valuable data, the Quantum Fireball Plus KA is for you, providing you're not annoyed by a louder than average hard drive.

Nortel Networks heeft op de Telecom99 beurs in Geneve een netwerk technologie gedemonstreert waarmee het 80Gbit/sec over een enkele golflengte weet te pompen. Da's een leuk draadje om een backbone van te breien:

Nortel showed 80-Gbit transmission over 480 km of dispersion-managed fiber manufactured by Corning, in a single span without regeneration. Anil Khatod, president of the optical Internet division at Nortel, said that in networks using carriers' existing single-mode fiber, transmission speeds of up to 40 Gbits/s per channel were possible. In new installations using dispersion-managed fiber, a full 80 Gbits/s per channel could be realized.

[...] Nortel simultaneously is expanding its optical packet-switching systems, supporting Internet Protocol over optical channels, with switching fabrics that start at 19.2 Tbits/s, scaling to hundreds of terabits. Nortel president John Roth said that more than 2,500 backbone IP routers could be reliably linked to the OpTera backbones as the new generation of fiber amps is rolled out.

Check EETimes voor het complete artikel.

De USB 2.0 Promotor Group (een groep van zeven bedrijven, waaronder Compaq, Hewlett-Packerd, Intel, Microsoft en Boonstra) hebben gisteren de specificaties van de USB 2.0 standaard bekend gemaakt. De bandwidth van het nieuwe USB touwtje zal naar verwachting 480Mbit/sec bedragen, da's 40 keer zo snel als de huidige USB standaard en lekker ook sneller dan Firewire . Hier heb je de press release, geleend van GA-Source:

A group of seven PC industry leaders announced today that the target speed of Universal Serial Bus (USB) 2.0 is 40 times faster than the existing USB 1.1. The achievement is anticipated to further build on the momentum of USB into more demanding user applications, such as image creation and interactive gaming.

The target speed of USB 2.0 is 480 Megabits per second (Mbs), as announced by the USB 2.0 Promoter Group, consisting of Compaq, Hewlett-Packard, Intel, Lucent, Microsoft, NEC and Philips. The target speed announcement coincides with the release of the USB 2.0 specification draft to industry developers here. The specification draft enables vendors to begin their product planning and development of USB 2.0 products.

"The revised and significantly higher target rate provides an effective upgrade path for today's USB peripherals. It is the result of engineering studies and test silicon that concluded that 480 Mbs can be achieved while still maintaining full compatibility with USB 1.1," said Jim Pappas, director of Intel's technology initiatives. The previous target speed range of between 360 and 480 Mbs was announced by the Promoter Group in August at the Intel Developer Forum.

Benefits of USB 2.0

The increased bandwidth of USB 2.0 opens the door for PC peripherals with more functionality, including faster broadband internet connections, higher resolution video conferencing cameras, next generation printers and scanners and fast external storage units. USB 2.0 also will make today's applications more productive. For example:

Consumers with digital cameras using USB 2.0 technology will be able to download a "roll" of digital film in seconds, compared to minutes on the earlier version of USB. Consumers can back up a gigabyte of data from their PC hard drive in less than a minute on USB 2.0 versus about a half-hour on USB 1.1. Scanners can create a high-resolution digital image in seconds on USB 2.0, versus minutes on USB 1.1.

Because USB 2.0 is an evolution of the existing USB 1.1 specification, it will be fully forward and backward compatible with current USB systems and peripherals. Even with the new speed target, USB 2.0 will work with existing cables and connectors.

"Compatibility brings added benefits," Pappas said. "Consumers can continue to use the same peripherals and cables that were purchased for USB 1.1, so their investments are protected. On the other side of the retail counter, peripheral vendors will see additional sales opportunities as they can sell new high-speed peripherals into the entire installed base of USB-capable systems."

What's Next

The USB 2.0 specification is expected to be finalized in the first quarter of 2000. Leading systems and peripherals are anticipated in the marketplace in the second half of 2000, according to the promoter group. More information is available at www.usb.org.

AGN Hardware heeft een vergelijking van vier harddisk coolers in elkaar geknutseld. Hier een lulverhaaltje over het nut van harddisk coolers:

Do you need a hard drive cooler? The answer is definite, it depends. It depends on what kind of drives you are using and sometimes how many you are using. If you are running a real hard drive, a SCSI drive, (send complaints to TheDude ) and we know who you are, the answer is yes you should have a cooler. The SCSI drives operate using a higher spindle speed and generate more heat. For those of you that refuse to come over to the dark side and see the light (several people are rereading that last sentence again) and are sticking with IDE you may not need a hard drive cooler. If you are using a 7200 rpm hard drive (IBM) then you should probably buy a hard drive cooler. If you are running a 5400 rpm hard drive and it's the only one in your case, save your money.

Nou, mijn 7200rpm en 5400rpm IBM drives worden absoluut niet heet, maar dat kan ook komen omdat ze pal voor twee (12cm en 8cm) intake fans hangen .

HardOCP heeft benchmarks gepost van een naar 800MHz overgeklokte Intel PIII Coppermine 700 op een Abit BE6-II moederplank met PC133 SDRAM. Helaas zijn er geen vergelijkende benchmarks met een Athlon, dus over het prestatieverschil tussen beide processors valt nog niet zoveel te zeggen.

Update: JC heeft een opmerking over de benchmarks geplaatst:

HardOCP put up some benchmarks of the Coppermine, it seems. Very impressive, at 800MHz on 133MHz bus, it manages to meet and beat slightly the 600MHz Athlon at 3D CPUMark. In other benchmarks, especially the memory related ones, Coppermine does real nice, probably matching (and overcoming) an equally clocked Athlon in some cases, but we have no side by side comparison here. Thanks to -- umm, forgot the name of the guy who sent me this. He's some sort of webmaster or something.

GA-Source vroeg Abit of er er nog plannen zijn voor een AMD Athlon plank. Zoals te verwachten viel kreeg met een super lame antwoord:

The main focus of ABIT are retail products. This means our main goal is always to offer top quality to our customers. This is the fundament of our success, and we are proud of it. Even though the capacities of the Athlon CPU are impressive, we need to consider also other factrs before we engage in the design of a new product. The current situation when looking at chipset and the entire situation indicates that the Athlon is at the moment more suitable for OEM products.

We are therefore regularely reviewing the situation, and will release a retail product as soon as we can be sure that our product will be able to satisfy completely the demands of our endusers.

Tja, als er één moederplanken fabrikant is die per sé Intel te vriend moet houden, dan is het wel Abit (dual Celerons bordjes e.d. zijn toch geen ideeën waarmee je je populair maakt bij Intel)...

Intel heeft de financiële cijfers van het derde kwartaal bekend gemaakt. Hoewel de omzet boven verwachting steeg, bleeft de winst achter. Yup, slechts 1,9 miljard dollar :

For the third quarter, Intel reported revenues of $7.33 billion, slightly more than the $7.3 billion expected. Earnings, however, came to $1.9 billion, or 55 cents a share, before accounting for acquisitions. While this represents an increase of 22 percent over earnings of 45 cents a share for the third quarter of 1998, it is less than was expected for the period. Analysts polled by First Call expected earnings of 57 cents a share, or nearly $1.97 billion. A number of analysts expected earnings to reach 59 cents per share.

[...] "Revenue growth was a little better than expected. Margin percentage was a little worse than expected," said Andy Bryant, Intel's CFO, on the conference call. Gross margin dipped from 59.4 percent to 58.7 percent, he said. Intel expected gross margin to rise.

The decline resulted from three factors, he said. The average selling price of Intel's PC chips fell because of an increase of low-cost PC sales combined with lost potential revenue from the delay of "some performance processors," Bryant said, or, in other words, the delay of the "Coppermine" Pentium III processors. Originally due in September, the Coppermine chips, which will run at 733-MHz, are now due on October 25.

In addition, Intel was hit with larger than expected costs associated with inventory and shifting to the more advanced 0.18-micron manufacturing process, which will be used with the Coppermine chips, he said.

In August, the company pushed back the release of the "Coppermine" Pentium III processors from an expected September release to late October because of problems in manufacturing the 600-MHz version of the chip. The delay also pushed back the release of Pentium III chips for notebooks.

Fiets richting News.com voor het complete verhaaltje.

Ik kwam via AMD Zone bij de site van Frosty Fox (de naam zeg het al), een echte die-hard die niet alleen de multiplier van z'n Athlon op via de 'harde' weg (weerstanden solderen) heeft aangepast, maar er ook nog eens een heftig waterkoel systeem tegenaan gooit. Dit alles maakte z'n Athlon 500 geschikt voor 700MHz, zeker geen slecht resultaat. Hier twee pics van z'n site:

Games Mania heeft een kort interview met Epic's Mark Rein op het net geprikt. Hier een vraag over de release dat van Unreal Tourney:

Gamesmania: Unreal Tournament must be getting close to completion. When will it be out?

Mark Rein: When it's done! On Epic's end it looks like we'll be done before the end of this month. How long it takes to get it in boxes and show up on store shelves is out of our hands.

Ik heb die demo trouwens afgelopen weekend gespeeld op een kleine LAN party en dat spel rouleert bruut de pan uit! Extreem heftige & snelle actie!

Ik lees bij Overclocker's Workbench dat MSI een nieuwe BIOS update voor z'n Athlon plank heeft klaar gezet:

MSI has releases their new Bios V1.3 for their popular Athlon motherboard. Those people that owned this motherboard has faced some problems in booting up their system with some PCI peripherals, can try this Bios revision. MSI Website: http://www.msi.com.tw.

MacOS Rumors heeft een flinke posting neer gemikt over de huidige stand van zaken met de ontwikkeling van MacOS X, de op Unix/Openstep gebaseerde opvolger van het (technisch) antieke MacOS 8.x. Het belangrijkste wat de Mac IMO nodig heeft is een stabiel OS en een paar kleine aanpassingen aan de interface (ik wil een taakbalk en een overzichtelijk startmenu, surfen op een Mac lukt me nog wel redelijk, maar echt werken met dat ding vind ik met het huidige MacOS ronduit frustrerend). Zolang Saint Steve de touwtjes strak in handen blijft houden zal-ie voor ons tweakers wat betreft hardware waarschijnlijk nooit interessant worden. Maareuh, terug naar OS X: de consumer release wordt voor begin volgend jaar verwacht, en zal de Mac eindelijk voorzien van preemptive multitasking, protected memory en allerlei andere blije features:

OS X as it stands today is already an industry-leading operating system. It brings to the Mac the preemptive multitasking, protected memory, distributed objects, clustering supporting, high-performance I/O, simple migration from Mac OS 9, powerful development environment, and many other features that the platform has lacked in the past.

However, its user experience is basically unchanged from the current Mac OS; the Finder includes some useful features borrowed from OpenStep's Workspace Manager and other subtle changes to dialogs and menus, but otherwise the experience is nearly identical to what today's Mac user works with in OS 8.6 and 9.0.

On first glance, this is a good thing; Apple has not taken away the advantages of the Finder's clear and simple interface. However, to make the introduction of OS X the major event that it must be for Apple to continue to grow, more must be done. Apple wants to add several unique new interface elements to the Finder, system dialogs, and menus that will revolutionize the industry; they also want to have a simple and well-understood driver/add-on system so that both hardware and software developers can begin supporting the OS right away.

As a result, Apple appears -- if the current TIME interview with Steve Jobs is any indication -- to have taken the more difficult, but also more rewarding, path by slowly preparing the Mac world for a "preview" rather than "final" Consumer Release early next year, two to three months after the Developer Preview 2 release. DP2 is now expected in very late '99 or at Macworld San Francisco 2000 in January.

This Consumer Preview Release (CPR?) will include the bulk of the interface improvements, performance optimizations, and driver/add-on support of the final release, but will be perhaps slightly less stable (a relatively minor issue, thanks to Protected Memory) and support a narrower range of hardware than the Final release, which is set for a Macworld New York (July 2000) deadline according to sources.

So, the news is certainly mixed -- but as long as steady progress continues on OS X, and Apple has made a clear committment to releasing it as soon as possible without compromising on its features, our outlook remains a positive one.

Check MacOS Rumors voor meer info.

Audiograbber is geloof ik wel een redelijk populair audio CD ripper. Betanews weet ons te vertellen dat er een nieuwe versie is gereleased:

After a few beta releases to registered users, a final version of Audiograbber 1.60 is out. Audiograbber is a beautiful piece of software that grabs digital audio from cd's. It copies the audio digitally- not through the soundcard- which enables you to make perfect copies of the originals. It can even perform a test to see that the copies really are perfect. Audiograbber can also automatically normalize the music, delete silence from the start and/or end of tracks, and send them to Fraunhofers acm codec or L3enc for automatic creation of MP3's. There are a lot more functions in Audiograbber, but to put it simply: Audiograbber has the most features one can wish from such a program.

Het proggie kun je hier downloaden. Verder bij Betanews:

- Creative Labs 3D Blaster Savage 4 Win9x drivers (2,6Mb)
- Asus Live met SECAM support beta 2 (1,3Mb)
- 3dfx Voodoo3 MacOS 8.x drivers 1.04 beta 4 (406Kb)

Michael van AnandTech heeft niet de primeur met de eerste review van een Riva TNT2 Pro kaartje. De Pro wordt gebakken op een nieuw 0,22micron fabricageproces, desondanks is de kloksnelheids slechts 143/166MHz (de TNT2 Ultra blijft dus gewoon sneller):

Immediately noticeable from the performance numbers is just how little the clock of a TNT2 affects the performance of a slower CPU such as the K6-2 450 or Celeron 300A, that are clearly CPU limited in these benchmarks. Furthermore, we can see that the memory bandwidth of the TNT2 Pro is somewhat limited at 166 MHz and this hinders its performance considerably. Overclocking to 170/180 puts the TNT2 Pro at or near the top in almost every benchmark. Otherwise, no surprises here as the TNT2 Pro is just a regular TNT2 with a higher clock speed.